Current Situation

We're experiencing performance challenges with the automatic refreshOidMap query that executes on every PostgreSQL connection. I'd like to hear from the community if others have faced similar issues and how they've approached it.

Our environment:

• ~1000 PostgreSQL schemas
• 200+ tables per schema
• AWS Lambda architecture (frequent connection establishment)
• The refreshOidMap query takes several seconds due to scanning all types across all schemas

🔍 The Query in Question

Currently, this query runs on every connection inside PostgresConnectionManager.#refreshOidMap():

WITH ranges AS (
  SELECT pg_range.rngtypid,
         pg_type.typname  AS rngtypname,
         pg_type.typarray AS rngtyparray,
         pg_range.rngsubtype
  FROM pg_range
  LEFT OUTER JOIN pg_type ON pg_type.oid = pg_range.rngtypid
)
SELECT pg_type.typname,
       pg_type.typtype,
       pg_type.oid,
       pg_type.typarray,
       ranges.rngtypname,
       ranges.rngtypid,
       ranges.rngtyparray
FROM pg_type
LEFT OUTER JOIN ranges ON pg_type.oid = ranges.rngsubtype
WHERE pg_type.typtype IN ('b', 'e');

This scans all PostgreSQL types system-wide, regardless of actual usage.

📊 Performance Impact Snapshot

🧠 Real Impact (pg_stat_statements Top Queries)

💡 Note: The first query is internal to Sequelize and accounts for the majority of total execution time. The remaining ones are application-level queries.

❓ Open Questions for the Community

• Has anyone faced similar performance issues with refreshOidMap in large multi-schema or serverless environments?
• Did you find any effective way to cache, override, or disable the automatic execution?
• Would the Sequelize team consider scoping the query by schema or making it configurable?
• Are there any Sequelize-specific strategies or hooks we can leverage to work around this?

Thanks in advance! 🙏